[00:00:01] Speaker 03: We have six cases on the calendar this morning, two from the PTAB, two from district courts, two from the MSPB, the latter being submitted on the briefs and not be argued. [00:00:19] Speaker 03: Our first case is BE Technology versus Twitter and Google, 2023, 11, 26, 27, 28, 30, 31, and 32. [00:00:30] Speaker 03: Mr. Neruzzi, before you speak, and please don't let the clock start yet, we just got a 28-J letter from you calling to our attention an action of a district court six months ago, which you obviously knew of then. [00:01:00] Speaker 03: If this is important to us, we shouldn't really have gotten it quite a while ago. [00:01:05] Speaker 03: We spent a lot of time preparing these cases and getting something that you had six months ago and sending it to us two days ago addressed to the clerk of court who hasn't been the clerk of court here for quite some time. [00:01:22] Speaker 03: That suggests a kind of [00:01:25] Speaker 03: Sloppiness, maybe, or inattentiveness. [00:01:27] Speaker 03: Do you have any explanation? [00:01:29] Speaker 01: Your Honor, on the clerk of court issue, I apologize. [00:01:32] Speaker 01: On the timing, I appeared as counsel personally myself within the past month as new counsel in the case. [00:01:43] Speaker 01: as to why prior counsel did not submit the district court's claim construction decision, I don't have a response to that. [00:01:50] Speaker 01: I will say, though, that Rule 28J is agnostic as to whether the supplemental authority benefits or detracts from a party's argument. [00:01:58] Speaker 01: It only matters that it is relevant new authority. [00:02:02] Speaker 01: And the decision also was not submitted by counsel for appellees. [00:02:05] Speaker 03: In any event, your recent counsel, you're off the hook. [00:02:13] Speaker 03: It wasn't your responsibility. [00:02:15] Speaker 03: All right, let us proceed. [00:02:17] Speaker 01: Thank you, Your Honor, and may it please the court. [00:02:21] Speaker 01: Your Honor, with respect to the unique identifier issue in claim one, the board made a key error of claim interpretation and then proceeded to make an error, an erroneous finding of unpatentability based on Guyot that rested on its erroneous claim understanding. [00:02:40] Speaker 01: The board conflated [00:02:42] Speaker 01: the notions of identification of a user with the notion of identification of the computer. [00:02:49] Speaker 01: And claim one expressly requires computer identification. [00:02:55] Speaker 01: So the claim language recites the unique identifier identifying the computer. [00:03:02] Speaker 01: Now what the board did is it said there is an embodiment in the specification in which user identification [00:03:10] Speaker 01: can also serve as computer identification. [00:03:14] Speaker 01: And then it proceeded to say, effectively, that therefore user identification always suffices to serve as computer identification. [00:03:22] Speaker 04: What does it matter if all of those devices all point to a specific computer and identify a computer? [00:03:33] Speaker 01: Judge Rainnick, I want to make sure I address your question correctly. [00:03:37] Speaker 01: Are you referring to what is Todd and Gruyo? [00:03:40] Speaker 01: I'm sorry, what? [00:03:41] Speaker 01: Are you referring to what is taught in the GUYO reference? [00:03:43] Speaker 01: Yes. [00:03:44] Speaker 01: Yes. [00:03:44] Speaker 04: And are you saying that in GUYO, the... No, I understand your argument to be that you can have many different types of devices or apparatuses that each have a specific user ID that don't point to a computer ID, right? [00:04:07] Speaker 01: No, that is not my argument. [00:04:09] Speaker 01: All right. [00:04:11] Speaker 01: Yes. [00:04:11] Speaker 01: Yes, your honor. [00:04:12] Speaker 01: Thank you. [00:04:12] Speaker 01: And I think this will hopefully elucidate the distinction because user ID or subscriber data, as the term is referred to in Guyo, is something like your Gmail login ID. [00:04:24] Speaker 01: It is your personal identifier that gets you into your account, and you can use that [00:04:30] Speaker 01: in GU-YO and in many instances of computer systems across any different computer, so long as you have access to the relevant software, whether it be web-based or... Okay, maybe I'm sorry I didn't cut you off. [00:04:42] Speaker 02: Maybe I'm misunderstanding. [00:04:43] Speaker 02: Are you arguing that there's a technical reason in the record that a unique identifier can't identify both the user and the computer? [00:04:54] Speaker 01: No, that's not what I'm arguing, Your Honor. [00:04:56] Speaker 02: OK, so technically they could. [00:04:58] Speaker 02: You're just arguing that the construction that allowed for that in this case is an error? [00:05:02] Speaker 01: No, that's not quite what I'm arguing, Your Honor. [00:05:05] Speaker 01: So thank you. [00:05:06] Speaker 01: And I hope to make this issue very clear. [00:05:09] Speaker 01: The claim requires computer identification. [00:05:13] Speaker 01: Computer identification is not one and the same as user identification. [00:05:19] Speaker 01: And that goes to Judge Raina's question. [00:05:20] Speaker 01: User identification is the person [00:05:23] Speaker 01: computer identification is the device. [00:05:25] Speaker 01: The claim requires computer identification through a unique identifier. [00:05:29] Speaker 01: Now it is possible and the board was correct in finding that in some instances someone's user ID could be [00:05:38] Speaker 01: designed by the system in a way that it not only identifies that user as a person, but also the device that they are coming from, right? [00:05:46] Speaker 01: But that is not necessarily always the case. [00:05:49] Speaker 02: But does that matter? [00:05:50] Speaker 02: I mean, maybe I'm misunderstanding the arguments, but why would they have to show that it is always the case? [00:05:57] Speaker 02: If it's sometimes the case, don't they prevail and don't you lose? [00:06:01] Speaker 02: I don't get what the relevance of always is. [00:06:05] Speaker 01: Because in GU-YO, it is never the case that the user ID or the subscriber data actually identifies the computer. [00:06:13] Speaker 01: What happens in GU-YO is that there is a unique proprietary identifier that's the computer identifier in GU-YO. [00:06:21] Speaker 01: And that tells the server that a copy of the client application has been installed on a given computer. [00:06:27] Speaker 01: However, that unique proprietary identifier is not used to identify the computer when a user logs in and interacts with the server. [00:06:37] Speaker 01: The user ID is associated. [00:06:39] Speaker 04: It may not have been designed for that purpose, but it [00:06:42] Speaker 04: does achieve that effect, doesn't it? [00:06:45] Speaker 01: No, Your Honor, it does not. [00:06:46] Speaker 01: What happens in GU-YO is that there's a database association that says, I'm going to just make up your user ID for me, Neruzi123. [00:06:55] Speaker 01: And in the GU-YO system, my user ID, Neruzi123, [00:06:59] Speaker 01: could be associated with my computer, Nuruzi PC, and there would be a record that this user has used this computer in the past, and this computer has a copy of the client software application that Gruyo teaches. [00:07:16] Speaker 01: However, what is missing in Gruyo and what the claims require is that when I am logged in as Nuruzi123, [00:07:24] Speaker 01: GU-YO has no idea which computer I'm actually coming from. [00:07:27] Speaker 01: It doesn't know the computer that is interacting with its server, whether it is that Neruzi PC that's associated from a past instance with my user ID, or whether I'm at the library here at the Federal Circuit, or whether I'm at the Kinko's down the street. [00:07:41] Speaker 01: It doesn't know that. [00:07:43] Speaker 01: And claim one requires this specific identification in the language. [00:07:47] Speaker 01: So there's separate claim language. [00:07:49] Speaker 01: that talks about an association and the board focused about on that association aspect, but the board ignored the import of the language that requires actual identification of the computer and instead what it said was that in our specification of the 440 patent, there is a teaching where user ID is specific enough that it also identifies the computer that the user is coming from. [00:08:14] Speaker 04: That's the point I was making with you earlier. [00:08:17] Speaker 04: That's exactly the point I was making. [00:08:19] Speaker 04: Tell me why the specification is incorrect. [00:08:22] Speaker 04: What's wrong with it? [00:08:24] Speaker 01: No, it is not incorrect. [00:08:25] Speaker 01: It's just a particular embodiment that is not replicated in GU-YO. [00:08:30] Speaker 01: So the specification teaches two embodiments on this point. [00:08:33] Speaker 01: One is the simple embodiment that we would all imagine, where you have a user ID that says who the user is, and you have a computer ID that says what the computer is. [00:08:40] Speaker 01: And that's the application ID. [00:08:42] Speaker 01: It's an ID that's associated with a copy of the application. [00:08:46] Speaker 01: And the board excluded that embodiment from the scope of the claims, which was clearly erroneous. [00:08:51] Speaker 01: Then there's another embodiment in the specification where it says the user ID can serve both purposes. [00:08:57] Speaker 01: It can tell you who the user is and it can tell you who the computer is, which computer it is. [00:09:02] Speaker 01: And for example, in my prior analogy, we could say, I have a user ID, Neruzi123, and I come from my computer, NeruziPC. [00:09:12] Speaker 01: But when both of these items of information go to the server, for instance, and the server appends all that together as Neruzi123-PC. [00:09:22] Speaker 01: And now this one identifier, this one unique identifier, [00:09:27] Speaker 01: contains both items of information. [00:09:29] Speaker 01: The specification of our patent says, sure, you can do it that way too. [00:09:33] Speaker 01: But GUYO doesn't have that. [00:09:35] Speaker 01: GUYO doesn't have either of the two embodiments that the specification teaches. [00:09:40] Speaker 02: I'm not clear about that. [00:09:41] Speaker 02: I mean, can you look at, I thought the board's decision finds otherwise about GUYO, pages 52 to 54. [00:09:48] Speaker 02: Sure. [00:09:49] Speaker 02: Am I wrong about that? [00:09:51] Speaker 02: So am I misreading what the board found, or are you just disagreeing with the board's finding? [00:09:57] Speaker 01: I am disagreeing with the board's findings and that they do not focus on and do not meet the relevant claim language. [00:10:04] Speaker 01: So in page 52, we can see that it says at the bottom, we have already discussed above that GUYO subscriber and subscriber system are associated, right? [00:10:16] Speaker 01: So everything you will see throughout the board's decision is about this association. [00:10:21] Speaker 01: But the problem reveals itself when we look at appendix 56, for instance, [00:10:28] Speaker 01: And the board recognizes an argument that we made near the top of page 56 that the computer and the subscriber are separate entities under the 440 patent. [00:10:46] Speaker 01: It recognizes that we made this point, but then it rejects that and it says, because the 440 specification discloses among other embodiments, that the user ID identifies the user [00:10:57] Speaker 01: dot, dot, dot and the computer, right? [00:11:00] Speaker 01: And that's one embodiment. [00:11:03] Speaker 01: And then from there the board says in the next sentence, the claim language and the specification do not support the contention that the user ID and the computer must each be separately identifiable. [00:11:13] Speaker 01: Well, that's incorrect. [00:11:15] Speaker 01: They each do need to be [00:11:17] Speaker 01: distinctly identifiable. [00:11:19] Speaker 01: It can be done by a single identifier in some instances, but that doesn't mean that just because GU-YO has a user ID that it serves to identify the computer. [00:11:27] Speaker 02: I'm trying to just sort out in my mind whether you're disagreeing with the construction that the board had or whether you're disagreeing with the application of its construction in terms of what the reference taught. [00:11:39] Speaker 02: In your first dialogue with me, it seemed like you agreed with the construction. [00:11:43] Speaker 02: You just thought the prior art didn't meet that construction. [00:11:46] Speaker 02: And now it seems like you're saying both. [00:11:50] Speaker 01: I'd like to clarify that. [00:11:52] Speaker 01: OK. [00:11:52] Speaker 01: So the board made errors at both levels. [00:11:54] Speaker 01: And its error in terms of the construction understanding led to its error in terms of the application of the prior art to the claims. [00:12:01] Speaker 01: And this is how those two errors unfolded. [00:12:04] Speaker 01: With respect to the understanding of the claim, the board [00:12:09] Speaker 01: looked to the language unique identifier throughout the claim and it also looked to the teachings around identifiers in the specification and it saw that in an embodiment, the unique identifier could be a user ID and it would meet the requirements of both identifying the user and the computer. [00:12:30] Speaker 01: then it went on to say, and this is a key error we haven't talked about yet, that- Wait, do you disagree with that or agree with that first point? [00:12:38] Speaker 01: That is one embodiment that is in the specification and it is correct that that is an embodiment and it would meet the claims because it does both jobs and namely the key job of identifying the computer. [00:12:50] Speaker 01: Right, so an embodiment in which... So it meets both, okay. [00:12:52] Speaker 01: In that embodiment. [00:12:54] Speaker 01: You agree. [00:12:54] Speaker 02: But you would have to... Well, can you answer my question? [00:12:56] Speaker 02: I mean, okay, it does it in the embodiment, so do you agree that that's a correct claim construction by the board? [00:13:03] Speaker 01: That part of it is not a claim construction. [00:13:06] Speaker 01: The claim construction problem I was going to address has not come up yet. [00:13:11] Speaker 01: So the board goes on to say that the language of the claim says selecting an advertisement to be displayed on the computer, the selection based at least on information associated with the unique identifier identifying the computer, and it says, well, [00:13:28] Speaker 01: It sort of stops reading at, based at least on information associated with a unique identifier. [00:13:34] Speaker 01: And it says the specification only teaches using the user ID to select an app. [00:13:40] Speaker 01: And even though the specification teaches a separate thing called an application ID that identifies the computer, that's not used to select an ad. [00:13:48] Speaker 01: Therefore, computer ID is not part of this claim. [00:13:52] Speaker 01: That's what the board said. [00:13:53] Speaker 01: And that's wrong because the claim language says the selection based at least on information associated with the unique identifier. [00:14:01] Speaker 01: User ID is associated with the application ID in the specification. [00:14:05] Speaker 02: Where did the board say the computer identifier is not in this claim? [00:14:08] Speaker 01: Yes, so you can look at, for example, Google's brief at 39. [00:14:14] Speaker 02: I'd like to look at the board opinion. [00:14:16] Speaker 01: Sure, appendix 25, appendix 62. [00:14:21] Speaker 01: Appendix 60 to 61, and you will see if you look at Google's red brief at 39 that they agree to the same thing, that that is what the board found. [00:14:31] Speaker 01: And so because the board thought that the concept of a computer identifier is not even within the scope, [00:14:38] Speaker 01: of claim one, and that only user ID can meet the claim, which is contrary to the specifications teachings, the board then did not require Ruyo to have any teaching that actually expressly identifies what computer is coming to the server. [00:14:54] Speaker 03: Counsel, you're into your rebuttal time. [00:14:57] Speaker 01: Yes, Your Honor. [00:14:58] Speaker 03: I assume you want to save it. [00:14:59] Speaker 01: I do. [00:15:00] Speaker 01: And I know I have not yet addressed cross-appellant's argument about real time. [00:15:05] Speaker 01: I'm happy to address that as the court would like. [00:15:09] Speaker 03: We'll hear from Mr. Cully. [00:15:12] Speaker 01: Thank you, Your Honor. [00:15:20] Speaker 00: Morning, Your Honors. [00:15:20] Speaker 00: May it please the Court. [00:15:23] Speaker 00: There is nothing in this specification or the claims of BP's patents that cast any doubt on the board's construction that a user identifier can also perform the functional requirements associated with the unique identifier recited in the claim [00:15:39] Speaker 00: In other words, being associated with the computer and identifying the computer. [00:15:44] Speaker 00: There was no exclusionary language in the claim to prevent that. [00:15:47] Speaker 00: The first time this court looked at this issue, and there's no more information to claim now. [00:15:53] Speaker 00: And for that reason alone, this court should affirm the board's claim construction. [00:15:57] Speaker 02: And your view is that the prior argument with the name, I can't pronounce, Gougeot, meets that? [00:16:05] Speaker 00: Yes, Your Honor, absolutely. [00:16:07] Speaker 00: So going on from the claim construction, so the claim construction dispute was most of what they briefed. [00:16:13] Speaker 00: And I don't think there's anything in the briefing, and I certainly didn't hear anything today that casts any doubt on how the board viewed the claim. [00:16:20] Speaker 00: So the next question is whether or not there is substantial evidence in the prior art that it does that. [00:16:26] Speaker 00: Now, what counsel just pointed this court to was a discussion on the bottom of page 52 of the board's decision. [00:16:34] Speaker 00: And he read the sentence that begins third, [00:16:36] Speaker 00: Third, we have already discussed above that GEO's subscriber and subscriber system are associated. [00:16:42] Speaker 00: And so council just made the argument that the board was fixated on this association. [00:16:48] Speaker 00: That yes, they're associated. [00:16:49] Speaker 00: The board kept saying they were associated. [00:16:51] Speaker 00: But that's not enough for it to be a identification of the computer. [00:16:56] Speaker 00: But what counsel did not read was the next sentence, which says, Dr. Hugh explains, and we agree, that with respect to the single user, single computer scenario, the subscriber's identification information stored in the subscriber data is the user identifier, and that because there is a subscriber system associated with the subscriber, the subscriber identification information is associated with the computer through its association with the subscriber. [00:17:20] Speaker 00: In other words, in the single user, single computer scenario, [00:17:25] Speaker 00: Once that identifier is associated with a computer, it necessarily identifies that computer, because there is only one computer, the computer the user is using. [00:17:35] Speaker 00: Now, what counsel did instead was he pointed the court to page 56. [00:17:39] Speaker 00: And page 56 is where the board discusses the testimony of their expert, Mr. Zakovich, which the board did not credit. [00:17:47] Speaker 00: So what he pointed you to was their expert's testimony. [00:17:51] Speaker 00: And their expert said what their expert said, but our expert said something else. [00:17:54] Speaker 00: The board weighed the evidence and didn't credit their expert. [00:17:57] Speaker 00: That's a simple, substantial evidence question. [00:18:00] Speaker 00: And he can point to his expert all he wants, but that doesn't mean that there wasn't substantial evidence in what our expert said and what was in the reference itself. [00:18:08] Speaker 00: So the single user, single computer embodiment gets to the identification of the computer once that association is made. [00:18:16] Speaker 00: And even if there wasn't a single user, single computer embodiment, the board explained that once a user logs on from a computer, at that point in the claim, there is only a single computer, and that's the computer the user logged on from. [00:18:33] Speaker 00: Now, the last thing I want to say about their appeal is that council pointed to the board's discussion of the unique installation identifier. [00:18:43] Speaker 00: And I believe he referred to page 44 of our red brief where we discussed that. [00:18:48] Speaker 00: That's not a computer identifier. [00:18:50] Speaker 00: That's the installation identifier. [00:18:52] Speaker 00: And the board said, well, that's not enough, because the installation identifier that you discuss in your patent has nothing to do with this targeted advertising. [00:19:00] Speaker 00: And the board was right about that. [00:19:03] Speaker 00: Moreover, the thing about the installation identifier is that it also is one step removed from the computer. [00:19:08] Speaker 00: It just identifies the installation, and through that, apparently they agree, identifies the computer. [00:19:13] Speaker 00: That's exactly what happens with a user ID. [00:19:16] Speaker 00: It identifies the computer because it's associated with the computer from which the user has logged in. [00:19:23] Speaker 00: Now, unless the court has further questions about their main appeal, I'll shift to the cross appeal. [00:19:32] Speaker 00: Now, just as in claim one, there's no exclusionary language in claim 25 to exclude processes that include any amount of consideration of previous user interaction. [00:19:46] Speaker 02: Well, that's true, because it says at least in part, right? [00:19:49] Speaker 00: It is true. [00:19:50] Speaker 02: But still, doesn't the prior reference have to include the part? [00:19:55] Speaker 02: that does what the claim says, the real time part, right? [00:20:00] Speaker 02: I mean, just because at least in part covers other things doesn't mean that the prior reference doesn't also have to cover what this is, the part that does it in real time. [00:20:12] Speaker 02: Unless I misunderstood the argument, I didn't understand that argument. [00:20:14] Speaker 02: I understood your claim construction argument, but I didn't understand that argument. [00:20:19] Speaker 00: So Your Honor, our argument is that once the board made that construction, [00:20:25] Speaker 00: that we believe clearly establishes the board's understanding that if there's any consideration of past user interaction, that process is beyond the scope of the claim. [00:20:35] Speaker 00: Once it made that construction, it then went to the prior art and just looked for some sort of use of past user interaction. [00:20:41] Speaker 00: And once it found it, it said, OK, that's not within the scope of the claim. [00:20:45] Speaker 00: And Your Honor. [00:20:45] Speaker 02: But does the prior art disclose at least some use of a real time? [00:20:52] Speaker 02: the at least in part that actually does it, that does the real time if the board construed it. [00:20:56] Speaker 02: That's what I was missing. [00:20:57] Speaker 02: It seemed like, yes, we know. [00:20:59] Speaker 02: Did you say that you have a prior art reference that does it? [00:21:04] Speaker 02: It just doesn't do it all the time? [00:21:07] Speaker 00: Yes, your honor. [00:21:08] Speaker 00: So both Gio and Robinson do it. [00:21:10] Speaker 00: Gio does it in the embodiment where they place an ad in response to a user action, like moving the mouse. [00:21:17] Speaker 00: And Robinson does it when it decides to put up a resonant ad that resonates with the web page that the user has gone to. [00:21:24] Speaker 02: But don't they fail because under the board's construction, real time doesn't include what you've already downloaded previously? [00:21:33] Speaker 02: Do these references have something that wasn't downloaded previously? [00:21:37] Speaker 00: They have a selection that's made in real time. [00:21:41] Speaker 00: Yes, Your Honor, and the board said that. [00:21:42] Speaker 00: They said that on page 97, the fourth line. [00:21:45] Speaker 00: It says, when Gio describes choosing and selecting advertisements, those actions are not performed in connection with reactive [00:21:52] Speaker 00: targeted advertising in real time. [00:21:55] Speaker 02: Rather, Gio's target- I'm sorry, I'm trying to follow you. [00:22:00] Speaker 00: So where are you on page 57? [00:22:02] Speaker 00: I'm on 97, fourth line down, towards the end, there's a sentence that begins, when Gio describes choosing and selecting advertisements. [00:22:09] Speaker 00: And that's referring to the selection that occurs upon user activity. [00:22:15] Speaker 00: And that's also at the bottom of the previous page. [00:22:18] Speaker 00: Where it says, and now. [00:22:20] Speaker 02: But where do they say that because it does that, but it doesn't do it all the time, that's the failure of Gujau. [00:22:26] Speaker 02: I couldn't glean that. [00:22:28] Speaker 00: OK. [00:22:29] Speaker 00: So it's not that it doesn't do it all the time. [00:22:31] Speaker 00: It's that when it does it, the selection is being made. [00:22:35] Speaker 00: from a group of ads that themselves were selected. [00:22:39] Speaker 02: No, that's what I mean. [00:22:40] Speaker 02: That's what I understood. [00:22:41] Speaker 02: The problem was it wasn't that you didn't meet the board's definition of real time. [00:22:46] Speaker 02: You just say that they misconstrued real time, and real time should include what was previously downloaded. [00:22:52] Speaker 00: Well, it shouldn't include what was previously downloaded. [00:22:54] Speaker 00: What it should include is a further selection from among the ads that were previously downloaded, which is exactly how their system works. [00:23:03] Speaker 00: So the board looked for [00:23:05] Speaker 00: this reliance on some amount of prior user interaction. [00:23:10] Speaker 00: And when it found it, it decided that consistent with its construction, which was wrong, that the reference was outside of the scope of the claim. [00:23:19] Speaker 00: And the problem with that is that they misunderstood the two tiers of advertising that are described in the patent. [00:23:26] Speaker 00: And they instead looked at them as two distinct types of advertising. [00:23:31] Speaker 00: So the way the two tiers work in their system, [00:23:34] Speaker 00: is first the system uses the demographic data to find a bunch of ads that would be targeted to the user. [00:23:40] Speaker 00: And then those ads are downloaded to the user's computer. [00:23:44] Speaker 00: Then, when the user's using the computer, the system detects some sort of user interaction and then selects from those downloaded ads an ad in response to that user interaction. [00:23:57] Speaker 00: That second part, the second tier, [00:24:00] Speaker 00: is where the real-time advertising that's recited in claim 25 comes in. [00:24:06] Speaker 02: See, that's what I'm misunderstanding. [00:24:08] Speaker 02: And I'll let you go after that. [00:24:09] Speaker 02: You can explain it. [00:24:10] Speaker 02: But I thought real-time under the board's construction did not include going to ads that were already downloaded, that it had to be a fresh thing, a new thing in real time. [00:24:22] Speaker 02: That's what I understood the disagreement was between you and the board, perhaps. [00:24:27] Speaker 00: So if the only thing that's happening is that an ad is presented without a further selection just because it was downloaded to the computer, that, the board would say, is simply relying on past user interaction. [00:24:41] Speaker 02: Yes. [00:24:41] Speaker 02: OK. [00:24:42] Speaker 02: Yeah. [00:24:42] Speaker 02: And you're saying that's wrong. [00:24:44] Speaker 00: We're saying that if there's a future selection, in other words, this group of ads is downloaded to the computer, okay, they're all based on previous user interaction. [00:24:53] Speaker 00: Now the question is, does the second tier kick in? [00:24:57] Speaker 00: Is there a further selection of an ad from that group based on user interaction? [00:25:03] Speaker 00: The board, although it seemed to understand that Geo and Robinson taught such an additional tier two action, a further selection, that it didn't matter because the selection was from among the ads that were already targeted. [00:25:19] Speaker 02: So it's as if the board was- So if we accept that, then the prior art reference does not teach what's in the claims. [00:25:24] Speaker 02: If we accept the board's view of that, which it can't include, at least in part, downloaded stuff that was previously downloaded, [00:25:32] Speaker 02: then the reference would not read on this, right? [00:25:35] Speaker 00: If you agree with the board that the claim excludes a third rail, if there's any use of that, yes, I think I'd have to agree with that. [00:25:44] Speaker 00: And Your Honor, not even BE agrees with that. [00:25:47] Speaker 00: And I would point the court to page 44 of the yellow brief in the middle, just above the middle, right after that site that says APPX. [00:26:00] Speaker 00: 30, a bunch of zeros in front of it. [00:26:02] Speaker 00: It says, the board never indicated that a process would not practice claim 25 if it involves first selecting a group of advertisements based on past user interaction, like demographic information, but then subsequently selecting from that group a targeted advertisement in real time. [00:26:21] Speaker 00: So Judge Prost, I understood your question just now to be, [00:26:26] Speaker 00: We agree with the board that what is outside the scope of claim 25 is any process that uses a previous user interaction to first select a subset of ads. [00:26:39] Speaker 00: Would that put the prior art outside the scope of the claim? [00:26:43] Speaker 00: Not even BE agrees that that is the case. [00:26:46] Speaker 02: That sentence seems to deal with this two tier thing, right? [00:26:50] Speaker 02: And not so much with the question I'm asking. [00:26:53] Speaker 00: Oh, I'm sorry. [00:26:53] Speaker 02: You know this case better than me, but that's the way I read what they said on page, whatever page you showed us. [00:26:59] Speaker 00: 44. [00:26:59] Speaker 00: 44. [00:27:00] Speaker 00: So the two tiers, it's like- Because they talked about the demographic information. [00:27:06] Speaker 00: Right, and so that's the initial targeting for the first tier. [00:27:09] Speaker 00: So you take all the ads that you could have, and then you take this demographic information, and you do your first tier targeting. [00:27:15] Speaker 00: And so you take all those ads, and you [00:27:18] Speaker 00: you bunch them together and put them on the computer. [00:27:21] Speaker 00: And the reason you do that, by the way, is they don't require internet access. [00:27:25] Speaker 00: And so once you do that, then there's a further step in tier two. [00:27:30] Speaker 00: And that further step is a selection from among those ads that's based on current user activity. [00:27:37] Speaker 00: Can I ask you one quick question? [00:27:38] Speaker 00: Yes. [00:27:38] Speaker 02: Because I'm not going to send you behind your time. [00:27:41] Speaker 02: I don't want to use it. [00:27:41] Speaker 02: Just one quick question. [00:27:42] Speaker 02: You're asking for reman. [00:27:43] Speaker 02: What would be done on reman? [00:27:45] Speaker 02: Is it an alternative to reversal? [00:27:47] Speaker 00: Okay, what would be done on remand is that the board would reconsider both Gio and Robinson with the correct construction. [00:27:53] Speaker 00: In other words, they wouldn't immediately dismiss those references as beyond the scope of the claim simply because they did their real-time selection from among a group of ads that were previously themselves selected. [00:28:06] Speaker 03: Thank you. [00:28:07] Speaker 00: And I'll reserve the remainder of my time. [00:28:17] Speaker 01: Thank you, Your Honors. [00:28:18] Speaker 01: I'd like to very briefly touch on a couple of points about the unique identifier and then move on to the real-time issue. [00:28:25] Speaker 01: Regarding the unique identifier point, counsel for cross appellants pointed to the bottom of page 52. [00:28:32] Speaker 01: I already addressed this issue. [00:28:35] Speaker 01: What that sentence talks about is an association between user ID and a past computer download of the application. [00:28:43] Speaker 01: It doesn't tell who you're. [00:28:45] Speaker 01: whether the user is using that computer right now. [00:28:48] Speaker 01: By contrast, if you look at appendix 395, which is our patent, column 29, line 65 to 67, you will see that our patent teaches that every single time the user comes to the server, [00:29:02] Speaker 01: the computer is also separately identified, or distinctly identified is what I mean by separately identified. [00:29:09] Speaker 01: Not just the user, but also the computer. [00:29:12] Speaker 01: And Guyot doesn't do that, and the board never found any evidence that Guyot does. [00:29:17] Speaker 01: With respect to the cross appeal and real time, Judge Prost, you're absolutely right that the prior art has to also do the real time part. [00:29:26] Speaker 01: And that is really what the board's decision is about. [00:29:30] Speaker 01: So you have to view the decision in the context of the arguments that were presented. [00:29:37] Speaker 01: Petitioners said that the claim can be met based entirely on ad selection that is tied to past activity. [00:29:48] Speaker 01: And the board said no. [00:29:50] Speaker 01: You cannot solely meet the claim based on past activity. [00:29:53] Speaker 01: And that's, for example, at Appendix 32, where the board's construction says that past activity cannot be the basis for the ad selection. [00:30:02] Speaker 01: Now, in Guyot and Robinson, as the board found, the only basis for ad selection [00:30:07] Speaker 01: is past user activity. [00:30:09] Speaker 01: There is no real-time component as to the ad selection. [00:30:14] Speaker 01: GUYO does look to whether the user is active based on its mouse movements or keyboard strokes to decide whether to show it an ad, but the decision and the sort of analysis of whether the user is active doesn't affect ad selection at all. [00:30:29] Speaker 01: And it's not simply based on the fact that the ads were downloaded before, [00:30:33] Speaker 01: It's even on the issue of which ad is shown to the user. [00:30:36] Speaker 01: Gruyo does nothing with that decision based on the user's current activity in real time. [00:30:42] Speaker 01: And neither does Robinson. [00:30:43] Speaker 01: Robinson is all about these affinity groups and determinations based on an algorithm of what someone will want to see based on what other people like them have wanted to see in the past. [00:30:54] Speaker 03: So I think I'm the unique identifier of the time available to you. [00:30:59] Speaker 03: And that is now none. [00:31:01] Speaker 04: Thank you, Your Honor. [00:31:02] Speaker 03: Thank you for your argument. [00:31:03] Speaker 03: Mr. Kelly has some time on the cross appeal. [00:31:07] Speaker 00: Thank you, Your Honor. [00:31:08] Speaker 00: So just to address the cross appeal, the two points about the prior art references, what council just discussed was a finding that the board never was able to make because of its own construction. [00:31:20] Speaker 00: The board's findings about the prior art and claim 25 were cabined by its construction that as long as it could find past user interaction, [00:31:29] Speaker 00: the references were outside the scope of the claim. [00:31:32] Speaker 00: DO teaches, and I don't think they can dispute, that an ad is selected when the mouse is moved from among the ads that were already downloaded. [00:31:41] Speaker 00: And when it comes to Robinson, same thing. [00:31:44] Speaker 00: An ad is selected when a user goes to a website. [00:31:48] Speaker 00: Now, just quickly about Robinson, the Robinson issue has to do with this issue called resonance. [00:31:54] Speaker 00: And resonance is a [00:31:55] Speaker 00: observation by the system that certain ads might be more likely to be clicked on based on the website. [00:32:03] Speaker 00: So you might go to websites about surfing or skiing, and there might be ads for snow pants or swimsuits. [00:32:11] Speaker 00: And that decision is made in advance, sure. [00:32:14] Speaker 00: It decides which types of those ads are resonant. [00:32:18] Speaker 00: But when the user actually goes to a website about skiing, the reason [00:32:24] Speaker 00: And this is my own example. [00:32:25] Speaker 00: The reason the system shows an ad for snow pants and not a swimsuit is because the user just went to that website. [00:32:34] Speaker 00: It selected an ad in real time based on the user's activity. [00:32:39] Speaker 00: Now, the reason why the board said, well, that's still outside the scope of Flame 25, [00:32:44] Speaker 00: is that the board reasoned that, ah, but those associations, those affinities with the two types of websites, those were predetermined. [00:32:53] Speaker 00: So in the past, the system decided which ads would resonate with one website and which ads would resonate in the second website. [00:33:00] Speaker 00: And because of that, there's no real-time selection. [00:33:03] Speaker 00: But what the board was not able to get to because of its construction is the reality that there is real-time selection. [00:33:10] Speaker 00: It's the selection of the ad that was previously determined to resonate with that website because the user landed on that website, which was an action of the user. [00:33:21] Speaker 00: And I have nothing further. [00:33:22] Speaker 00: Thank you. [00:33:23] Speaker 03: Thank you, counsel. [00:33:24] Speaker 03: Thank you, both counsel, in case you submit.